¹Department of Cellular and Molecular Medicine, University of Ottawa, Ottawa Regional Cancer Centre, 501 Smyth Road, Ottawa, Ontario, Canada K1H 8L6

²Department of Medicine, University of Ottawa, Ottawa Regional Cancer Centre, 501 Smyth Road, Ottawa, Ontario, Canada K1H 8L6

³Department of Obstetrics and Gynaecology, University of Ottawa, Ottawa Regional Cancer Centre, 501 Smyth Road, Ottawa, Ontario, Canada K1H 8L6

Correspondence to: Barbara C Vanderhyden, Ottawa Regional Cancer Center, 501 Smyth Road, Ottawa, Ontario, Canada K1H 8L6

In preparation for ovulation, paracrine communication between the preovulatory follicle and overlying theca/stromal cells and ovarian surface epithelium (OSE) must take place to facilitate the degradative and apoptotic events associated with ovulation. Kit tyrosine kinase receptors and their ligand, kit ligand (KL) are expressed within ovarian follicles, and ligand-induced receptor activation appears to account for some of the cell - cell interactions important for oocyte development. We investigated the expression of Kit receptors and KL in OSE cells and the possibility that modulation of their expression could affect OSE cell activity. KL mRNA and protein were detected in the OSE cell layer of rat ovaries, and primary cultures of rat OSE as well as the immortalized rat OSE cell line, ROSE 199, expressed KL, but not Kit receptors. Both primary and immortalized OSE cells preferentially expressed KL-1, rather than KL-2, transcripts, suggesting that these cells produce predominantly the soluble form of KL. Activation of the cAMP signalling pathway using dibutyryl cAMP decreased proliferation of ROSE 199 cells and elicited a threefold increase in KL expression. TGF- similarly inhibited ROSE 199 cell proliferation, but strongly inhibited dibutyryl cAMP-induced KL expression, indicating that changes in KL expression were not directly associated with OSE cell proliferation. The expression of mostly soluble KL in the surface epithelium suggests that this cytokine may be acting in a paracrine fashion, perhaps interacting with nearby Kit receptor-bearing theca cells.

ovary; receptor tyrosine kinase; ovulation; cytokine; cell - cell communication

Introduction

The ovarian surface epithelium is a single continuous layer of epithelial cells that surrounds the ovary and is embryologically derived from the coelomic epithelium which overlies the genital ridge (Julian, 1974). In adult life, the ovarian surface epithelial (OSE) cells vary in phenotype from squamous to columnar (Nicosia et al., 1985), and contribute to the process of lysis and repair of the ovulatory site by secreting proteases that aid in follicular rupture (Bjersing and Cajander, 1975). Prior to ovulation, the OSE cells overlying the stigma undergo apoptotic cell death (Murdoch, 1996). After ovulation, OSE cells at the perimeter of the wound undergo a burst of proliferative activity to repair the ovulatory wound and thus restore the continuity of the epithelium (Murdoch, 1996).

With the exception of proteases secreted at the time of ovulation, little is known about OSE cells and their ability to regulate ovarian function. However, knowledge of OSE cell products that may regulate stromal cell function is of particular importance since OSE cells can form invaginations and cysts that give rise to ovarian epithelial carcinomas. OSE cells are capable of producing extracellular matrix proteins (Kruk and Auersperg, 1994), the cytokines IL-1, IL-6 and macrophage colony-stimulating factor (Lidor et al., 1993; Ziltener et al., 1993) and other growth-regulatory molecules, including TGF- (Jindal et al., 1994) and TGF- (Berchuck et al., 1992). Because of the potential relevance to both normal ovarian function and the development of ovarian tumours, further examination of OSE secretory products is warranted.

Both hormones and growth factors have been shown to influence OSE cell proliferation. Gonadotropins stimulate growth of normal rabbit OSE cells (Nicosia et al., 1985). Both rat (Godwin et al., 1992) and human (Rodriquez et al., 1991) OSE cells proliferate in response to EGF and TGF-, and EGF receptors have been identified in human OSE cells (Jindal et al., 1994; Berchuck et al., 1991). Although TGF- receptors have yet to be found on human OSE cells, TGF- has been shown to inhibit the proliferation of these cells in culture (Berchuck et al., 1992) implying the presence of the receptor for this growth factor. Specifically which of these factors are involved in the apoptotic and proliferative activities in OSE cells around the time of ovulation is not clear.

The Kit tyrosine kinase receptor and its ligand, KL, have been implicated in the growth regulation of a variety of cells. In fact, in most Kit-bearing cell types that have been examined, activation of Kit receptors by KL stimulates cell proliferation. KL stimulates colony formation of primitive hematopoietic precursors from bone marrow (Metcalf and Nicola, 1991) and induces the proliferation of mast cells (Tsai et al., 1991). Like somatic cells, primordial germ cells, which express Kit receptors, respond to KL in culture with increased survival (Dolci et al., 1991; Godin et al., 1991; Pesce et al., 1993) and proliferation (Matsui et al., 1991). In prepubertal male mice, KL selectively stimulates DNA synthesis in type A spermatogonia (Rossi et al., 1993).

Strong evidence for the involvement of Kit receptors in the regulation of OSE cell growth is provided by mice with mutations at the Kit locus. The Kit^x/v mutation is a substituted amino acid in the Kit receptor kinase domain resulting in reduced kinase activity (Nocka et al., 1990; Reith et al., 1990). In Kit^x/v mice, the surface epithelium of neonatal animals exhibited foci of crowded OSE cells up to several layers thick compared to non-mutant animals of the same strain (Murphy, 1972). This overgrowth ultimately led to the formation of extensive invagination of epithelial tubules into the ovary cortex and, by 13 - 20 weeks, these animals exhibited complex formations of tubular adenomas composed of invasive epithelial tubules and interstitial cells. Although the mechanism by which stromal-epithelial interactions are disrupted is currently unknown, the evidence that theca cells express Kit receptors (Motro and Bernstein, 1993; Manova et al., 1993) would suggest that theca cell function has been compromised by Kit^x/v mutations. The phenotype of the mutant mice would suggest that theca cells normally mediate growth inhibition of OSE cells since loss of normal Kit function results in OSE cell invasion into stroma.

Since Kit and KL interactions stimulate proliferation of a variety of cell types, we examined whether Kit and/or KL are produced in rat OSE cells and if their expression could be modulated. Given the limited information about the factors that affect OSE cell activity, particularly around the crucial time of ovulation, we investigated the effects of TGF-, a product of theca cells that is upregulated at ovulation, on both OSE cell proliferation and KL expression.

Results

Expression of KL and Kit in OSE cells

In situ hybridization was performed to localize KL mRNA transcripts in sections of 28-day-old rat ovaries. Parts of ovary sections from PMSG-primed rats are shown in Figure 1. Probe hybridization is evident as a dark blue-coloured reaction product. Staining was seen throughout the ovarian epithelium (Figure 1a,b) and was more evident at higher magnifications (Figure 1c). When sections of ovaries from PMSG-primed rats were incubated with excess unlabelled probe (Figure 1d), no non-specific binding was seen in the epithelium except in the hilium region.

To determine the presence of KL protein in rat OSE cells, sections of PMSG-primed ovaries were exposed to antibodies against KL and positive staining was seen as a diffuse dark brown reaction product. KL protein was detected throughout the surface epithelium (Figure 2a). When antibodies were omitted, the dark reaction product was not evident (Figure 2b) indicating that the reaction product was present only under circumstance of specific antibody-antigen reaction.

To further investigate the expression of KL and Kit in OSE cells, RNA was isolated from OSE cells derived from ovary explants and from ROSE 199 cells, and the presence of KL transcripts was detected by Northern blot analysis. KL transcripts were present in OSE and ROSE 199 cells (Figure 3a); however, these cells did not express message for Kit receptors (Figure 3b) as seen by a lack of signal in these preparations but an abundance of signal in the positive control (MC/9) lane.

The results of the in situ hybridization and northern blot analyses indicated that KL mRNA was present in both normal OSE and ROSE 199 cells. To determine which KL transcripts are produced by OSE cells, RT - PCR was performed using total RNA preparations from ROSE 199 cells and from two independently-established primary cultures of OSE cells. The PCR primers spanned the alternatively spliced exon and enabled the detection of both KL-1 and KL-2 transcripts (Huang et al., 1992). Although both KL-1 and KL-2 cDNAs were amplified in OSE and ROSE 199 cells, there was a predominance of KL-1 cDNA (Figure 4). Similar results were obtained with a second pair of PCR primers (data not shown). Thus OSE cells, both normal and immortalized, produced KL transcripts and the majority of these transcripts were in the form that, when translated, yield a protein that is preferentially cleaved to produce soluble KL.

Effect of dbcAMP and TGF- on expression of KL and Kit mRNA

Since no Kit was detectable in normal and ROSE 199 cells, and since Kit expression can be upregulated in other cell types by cAMP (Nishina et al., 1992; Ogawa et al., 1995), we investigated the possibility that activation of the cAMP pathway could induce Kit expression in these cells. However, no Kit transcripts were detected in cells that had been cultured in the presence of 1 mM dbcAMP, despite a strong positive signal in mast cells which served as a positive control (data not shown).

To determine if dbcAMP or TGF- could modulate the expression of KL in ovarian epithelial cells, ROSE 199 cells were treated for 48 h in the absence or presence of dbcAMP and/or TGF-. KL mRNA expression increased 3.0±0.2-fold when ROSE 199 cells were treated with 1 mM dbcAMP (Figure 5). In contrast, TGF- reduced basal expression of KL to levels approximately half that of untreated cells and was particularly effective in suppressing the dbcAMP-induced expression of KL sixfold, resulting in KL levels that were lower than those seen in untreated ROSE 199 cells (Figure 5). Despite the differing effects of dbcAMP and TGF- on the steady-state levels of KL mRNA, both treatments, as well as the combined treatment, resulted in similar proportions of KL-1 : KL-2, in all cases yielding a relative abundance of KL-1 (Figure 6).

Effects of dbcAMP and TGF- on ROSE 199 cell proliferation

ROSE 199 cells displayed a sevenfold increase in the average number of cells after 48 h in culture, suggesting a doubling time of approximately 16 h. Both dbcAMP and TGF- significantly inhibited the proliferation of ROSE 199 cells during a 48 h culture; the combined treatment was not more effective than the effects of dbcAMP alone. The inhibition of cell proliferation was not apparently associated with increased cell death, as >99% of the cells remained attached in all treatments, and >98% were viable, as indicated by the exclusion of Trypan blue.

Discussion

In this study we determined that KL is expressed in the surface epithelium of rat ovaries, in primary cultures of ROSE cells, and in immortalized rat OSE cells. DbcAMP increased the expression of KL in ROSE 199 cells, whereas TGF- inhibited both basal and dbcAMP-stimulated KL expression. Despite their opposing effects on KL expression, both dbcAMP and TGF- strongly inhibited ROSE 199 cell proliferation. Kit mRNA was not expressed in either normal or immortalized OSE cells.

Although this is the first demonstration of the presence of KL mRNA and protein in OSE cells of adult ovaries, KL mRNA has been reported in the surface epithelium of ovine ovaries during gestation (Tisdall et al., 1997). This expression in ovine OSE was lost at birth, however, so that no KL expression was seen in the postnatal ovary. Primary cultures of rat OSE cells, established from ovary explants, retained the expression of KL through several passages. Isolation of rat OSE from ovary explants yielded fairly homogeneous populations and one could readily distinguish OSE cells from other ovarian cell types based on differences in morphology of the cells (Adams and Auersperg, 1981). Rat OSE cells in culture had clear cytoplasm, well-defined borders and formed confluent monolayers, similar to those described by Adams and Auersperg (1981). Furthermore, the ovarian cells underlying the OSE cells are stroma cells, which express Kit (Manova et al., 1993; Motro and Bernstein, 1993; Lammie et al, 1994), so the absence of Kit expression in the rat OSE cell preparation would suggest that there was no stromal cell contamination in these cultures.

Both normal and immortalized OSE cells were shown to have a predominance of KL-1 mRNA which, when translated, would preferentially yield soluble KL. The expression of Kit receptors in the underlying theca cells in vivo suggests that the soluble ligand activates the theca cell Kit receptors to elicit a response. The responsiveness of theca cell Kit receptors to soluble KL has been recently demonstrated by Parrott and Skinner (1997) who reported that KL stimulated gonadotropin-independent theca cell proliferation as well as androgen production. The importance of Kit-KL interactions has been demonstrated previously in paracrine communication between oocytes and granulosa cells (Ismail et al., 1996; Packer et al., 1994), and it would seem that some aspects of theca-OSE cell communication may also be mediated by KL activation of Kit receptors.

Kit transcripts were absent in both primary and immortalized rat OSE cells and, although Kit expression in other ovarian cell types has been shown to be regulated by various factors including cAMP (Nishina et al., 1992; Ogawa et al., 1995) and human chorionic gonadotropin (Horie et al., 1991), it could not be induced in OSE cells by dbcAMP. Although this observation suggests that Kit does not play a direct role in OSE cell activity, it is equally possible that Kit is important, but is expressed in OSE cells only under specific conditions which were not duplicated in culture, or that Kit expression cannot be induced in the immortalized cells used in this study.

Gonadotropins, both follicle-stimulating hormone (FSH) and luteinizing hormone (LH), show elevated levels at the time of ovulation and have regulatory effects on gene expression in ovarian cells that have appropriate receptors. Although FSH has not been shown to regulate Kit expression, the LH-like hormone hCG decreases Kit expression in both theca cells and oocytes of mouse ovaries (Motro and Bernstein, 1993; Horie et al., 1991). In addition, both FSH and hCG-like hormones increase KL mRNA expression in granulosa cells of mice (Motro and Bernstein, 1993) and rats (Ismail et al., 1996). Although there is evidence that OSE cells express receptors for both LH (Osterholzer et al., 1985) and FSH (Zheng et al., 1996), the immortalized OSE cell line used in this study does not express these receptors (manuscript in preparation). However, cAMP serves as the second messenger following receptor activation by both gonadotropins, and activation of this signalling pathway can be mimicked by dbcAMP. Although Kit expression can be upregulated by activation of this pathway in some cell types (Nishina et al., 1992; Ogawa et al., 1995), Kit was not expressed in ROSE 199 cells treated with dbcAMP. In contrast, dbcAMP increased the expression of KL mRNA which is in agreement with its effects on KL expression in granulosa cells (Manova et al., 1993) and Sertoli cells (Rossi et al., 1991). Coincident with the enhanced expression of KL, dbcAMP also strongly inhibited ROSE 199 cell proliferation; however the lack of Kit receptors in these cells would suggest that the altered gene expression was not responsible for the changes in proliferation. While there is the possibility that the suppression of proliferation directly or indirectly caused the upregulation of KL expression, it seems unlikely as the growth-inhibitory effects of TGF- elicited opposite effects on KL expression. Therefore, it is reasonable to suspect that these two responses to activation of the cAMP signalling pathway are functionally unrelated.

The differential effects of cAMP on mitogenic signalling pathways is well established. In some cells, it promotes growth (Cass and Meinkoth, 1998), while in many others, it inhibits proliferation, both basal (Chen et al., 1998; Tortora et al., 1997), and growth factor-stimulated (Tortora et al., 1997; Vadiveloo et al., 1996; D'Angelo et al., 1997), through a pathway that includes both PKA and Map kinase (Chen et al., 1998; Tortora et al., 1997). Since OSE cells in vivo are generally quiescent with low proliferation rates associated primarily with wound repair after ovulation (Murdoch, 1996), it can be speculated that the cAMP signalling pathway contributes to the maintenance of the quiescent state of these cells by suppressing growth-promoting mechanisms.

Stromal and theca-interstitial cells of both human and bovine origin have been shown to produce soluble products that alter OSE cell proliferation (Vigne et al., 1994; Karlan et al., 1995). One key negative regulator is TGF- which is produced in increasing abundance by the theca cells during follicle growth (Skinner et al., 1987; Chegini and Flanders, 1992). Our results are in agreement with previous observations of the growth inhibitory effects of TGF- on OSE cell proliferation (Berchuck et al., 1992; Vigne et al., 1994), and suggest that theca-derived TGF- may elicit responses not only from the follicular granulosa cells (Skinner et al., 1987), but from the overlying OSE cells as well. In addition to its inhibitory effects on OSE cell proliferation, TGF- suppressed KL mRNA expression, which is in agreement with previous studies that demonstrated inhibitory effects on KL mRNA expression and activity in other cell types (Heinrich et al., 1995; Mekori and Metcalfe, 1994; Ramsfjell et al., 1997). Interestingly, TGF- not only reduced basal KL expression in OSE cells, but eliminated the dbcAMP-induced increase in KL expression, suggesting that activation of TGF- receptors and/or a downstream effector molecule can block cAMP-induced transcription.

Previous work published by our laboratory (Ismail et al., 1997) and others (Allard et al., 1996) have shown that hormonal and pharmacological stimuli can preferentially alter levels of expression of KL splice variants. However, despite their effects on KL transcription, neither dbcAMP nor TGF- elicited any changes in the proportion of the two KL transcripts generated in ROSE 199 cells, indicating that overall transcription levels, but not the mechanisms controlling alternative splicing, are regulated by these factors. The ability to regulate levels of KL production in OSE cells, and preferential generation of the transcript that encodes soluble KL, underscores the possibility that this ligand may act in a paracrine fashion on neighbouring cells that bear Kit receptors, most likely theca cells. As our knowledge of the mechanisms of communication between these two cell types improves, it will be possible to evaluate their contribution to the processes of ovulation and the formation of invaginations and cysts that give rise to ovarian epithelial carcinomas.

Materials and methods

Animals

Prepubertal and adult (3 - 12 month old) female Sprague-Dawley rats were obtained from the animal colony at the University of Ottawa. Animals were allowed free access to food and water, and lighting was provided for 14 h daily. Some prepubertal animals were injected intraperitoneally at 26 days of age with 7.5 IU of pregnant mare's serum gonadotropin (PMSG; Folligon, Intervet, Whitby, ON). Gonadotropin-treated animals were sacrificed 40 h after PMSG injection. Untreated prepubertal animals were sacrificed and used as age-matched controls. Adult rats were sacrificed, the ovaries dissected and used to isolate OSE cells.

Primary and immortalized OSE cells

OSE cells were isolated from ovary explants as previously described in the literature (Adams and Auersperg, 1981) with the exception that explants were allowed to dry for 7 min instead of 5 min. Explants were incubated in 5% CO₂: 5% O₂: 90% N₂ for 4 days in Waymouth MB 752/1 medium (Sigma, St. Louis, MO, USA) supplemented with 25% FBS (Hyclone) 100 IU/ml penicillin G (a generous gift from Sigma) and 100 g/ml of streptomycin (Sigma), with a media change after 3 days. After this period, a ring of epithelial cells was evident around each explant. The explant was removed and any contaminating cells were scraped using a 25 G needle. Cells were washed twice with Hanks' Balanced Salt Solution (Gibco - BRL, Burlington, ON) and then removed from the dish by exposing cells to 0.25% trypsin in phosphate-buffered saline (PBS) for 20 min at room temperature. OSE cells were plated at 5000 cells/ml in media and after two passages, the cells were harvested for total RNA. Six individual OSE cultures were established and used for RNA extraction and subsequent Northern blot analysis.

ROSE 199 cells, a spontaneously immortalized rat OSE cell-derived cell line (a generous gift from Dr N Auersperg; Adams and Auersperg, 1985), were used in studies investigating OSE cell proliferation, and Kit and KL mRNA expression. These cells were maintained in -minimum essential medium (Gibco - BRL) supplemented with 10% FBS in tissue culture dishes (Falcon), incubated 37°C in 5% CO₂.

In situ hybridization

Sections of ovaries from untreated prepubertal rats and age-matched PMSG-primed animals were prepared and used for in situ hybridization as described previously (Ismail et al., 1996). Briefly, KL mRNA was detected with digoxigenin-labelled oligonucleotide probes, whose sequence was complementary to the RNA encoding a portion of the extracellular region of KL. The sections were exposed to alkaline phosphatase-conjugated antibodies against digoxigenin and then incubated with alkaline phosphatase colour reaction substrates. A dark blue reaction product indicated positive KL mRNA signal.

Specificity of the probe was evaluated using the following controls: (1) omission of labelled KL probe from the hybridization incubation; (2) incubation using labelled KL probe with unlabelled probe in excess (100 : 1 unlabelled:labelled probe); (3) incubation with a heterologous probe complementary to nucleotides 151 - 174 of bacterial neomycin transferase and which does not share any common nucleotide sequences with KL. For all post-hybridization steps, control sections were developed according to the above protocol and together with sections which were hybridized with labelled probe. Endogenous alkaline phosphatase activity was assessed by omitting the digoxigenin antibody from the immunological detection procedure.

Immunohistochemistry

Paraffin-embedded sections of rat ovaries were cleared in a graded xylene/ethanol series and used for immunohistochemistry. Endogenous peroxide activity was blocked using 1% H₂O₂ (BDH) diluted in Stockholm's phosphate buffered saline (S-PBS). Slides were rinsed for 5 min with three changes in S-PBS, and the antigenic sites were exposed by treating slides with 0.1% trypsin in S-PBS for 15 min at room temperature. Slides were washed for 5 min with three changes in S-PBS and then incubated for 1 h at room temperature in a dark humidified chamber with 20 mg/ml of polyclonal goat anti-mouse KL neutralizing antibodies (R&D Systems, Minneapolis, MN, USA), diluted in diluent (0.3% triton X-100 in S-PBS). After incubation, slides were rinsed twice and washed three times for 5 min each in S-PBS. Sections were exposed to a diluted (1 : 200) biotinylated donkey anti-goat antibody for 30 min and washed as above. Slides were incubated with a horseradish peroxidase-conjugated Streptavidin antibody diluted 1 : 200 in S-PBS and again washed as described above. To determine reactive antigenic sites, sections were exposed using a diaminobenzidene substrate detection kit (Boehringer Mannheim, Laval, PQ) according to manufacturer's instructions. After colour development, slides were dehydrated and mounted with permount (Fisher Scientific, Ottawa, ON).

Northern blot analysis of Kit and KL mRNA expression

To determine the levels of mRNA expression of Kit receptor and KL in OSE and ROSE 199 cells, total RNA was isolated using either an RNeasy kit (Qiagen, Chatsworth, CA, USA), or a LiCl precipitation method according to a standard protocol (Sambrook et al., 1989), depending on the number of cells available. Total RNA was also extracted from MC/9 (ATCC, Rockville, Maryland, USA), NIH3T3 (ATCC) and Sl⁴ (generous gift from Dr D Williams) cells to serve as Kit positive/KL negative, KL positive/Kit negative controls, and KL negative/Kit negative, respectively. RNAs were separated on 0.9% agarose formaldehyde gels and transferred overnight to Hybond-N nylon membranes (Amersham, Oakville, ON) or nitrocellulose membranes (MSI, Westborough, MA, USA). The blots were hybridized with ³²P-cDNA probes labelled by random priming to either murine or human KL cDNA (gift from AMGEN, Thousand Oakes, CA, USA), or a 680 bp fragment isolated from PCR amplified rat KL-1 cDNA. These same blots were stripped and reprobed with ³²P-labelled 3 kb EcoRI - HindIII fragment isolated from mouse Kit cDNA (generously donated by Dr A Bernstein). To serve as a loading control, blots were probed a third time with ³²P-labelled probes for -tubulin mRNA (a gift from Dr McBurney). Intensity of bands were estimated using a PSI phosphoimager (Molecular Dynamics, Sunnyvale, CA, USA) and the ratios of Kit or KL to -tubulin mRNA were calculated and standardized to control values which were arbitrarily set to 1.0. Northern blots to determine levels of Kit or KL mRNA were done four times with six different samples of OSE cells and three samples of ROSE 199 cells. Northern blots to determine regulation of Kit and KL expression were done three times with three independent samples of each treatment group.

Differentiation of alternative KL transcripts

Reverse transcription-polymerase chain reaction (RT - PCR) was used to determine the relative proportion of KL-1 and KL-2 transcripts in normal and ROSE 199 cell total RNA preparations using a procedure and primers as described previously by our laboratory (Ismail et al., 1997). PCR products were then separated on a 1% NuSieve, 1% agarose gel. Some PCR reactions were done using water instead of RT sample to serve as a negative control. There were two positive controls; (1) C13, an epithelial cell line that produces predominantly KL-1 transcripts and; (2) granulosa cells that produce predominantly KL-2 transcripts (Ismail et al., 1997). This experiment was performed twice with two separate RNA preparations. In some cases, results were confirmed with a second pair of KL primers.

ROSE 199 cell proliferation

ROSE 199 cells were seeded at 1.5´10⁵ cells in 100 mm tissue culture dishes and incubated at 37°C for 48 h in 10 ml of -minimum essential media+10% FBS, in the presence or absence of 1 mM dibutyryl cyclic adenosine monophosphate (dbcAMP: Boehringer Mannheim), a membrane permeable analogue of cAMP and/or 10 ng/ml TGF- (R&D Systems). At the end of the culture period, plates were coded by an independent observer to allow for unbiased analysis and cells were assessed for morphological changes using a light microscope. Cells were subsequently detached from the dishes using 0.25% trypsin in PBS, transferred to 15 ml Falcon tubes, pelleted by centrifugation at 3000 g for 4 min, resuspended in 1 ml of media and counted on a hemocytometer. In some experiments, Trypan blue dye exclusion was used to estimate cell viability. After counting all samples, duplicate samples were pooled, and cells were pelleted, flash frozen in liquid nitrogen, and stored at -80°C until Northern blot or RT - PCR analysis. Experiments were performed three times in duplicate. Statistical comparisons were made using analysis of variance. When significant effects were observed, the Bonferroni post-test for multiple comparisons was used. Data are presented as the mean±s.e.m. and statistical significance was inferred at P<0.05.

Acknowledgements

The authors wish to acknowledge technical assistance with the OSE PCR provided by Megan Hersh. We wish to thank Dr N Auersperg for the gift of ROSE 199 cells and Dr D Williams for the Sl⁴ cells. We are also grateful to AMGEN for providing the PCR primers and rat KL cDNAs, and to Dr A Bernstein for the mouse Kit cDNA. We would like to thank Sigma for their gift of penicillin-G. This research was supported by funds provided by the Cancer Research Society and the Medical Research Council of Canada.

Adams AT and Auersperg N. (1981). Cancer Res. 41, 2063-2072. MEDLINE

Adams AT and Auersperg N. (1985). Exp. Cell Biol. 53, 181-188. MEDLINE

Allard EK, Blanchard KT and Boekelheide K. (1996). Biol. Reprod. 55, 185-193. MEDLINE

Berchuck A, Rodriguez GC, Kamel A, Dodge RK, Soper JT, Clarke-Pearson DL and Bast Jr RC. (1991). Am. J. Obstet. Gynecol. 164, 669-674. MEDLINE

Berchuck A, Rodriguez G, Olt G, Whitaker R, Boente MP, Arrick BA, Clarke-Pearson DL and Bast Jr RC. (1992). Am. J. Obstet. Gynecol. 166, 676-684. MEDLINE

Bjersing L and Cajander S. (1975). Experientia 31, 605-608. MEDLINE

Cass LA and Meinkoth JL. (1998). Endocrinology 139, 1991-1998. MEDLINE

Chegini N and Flanders KC. (1992). Endocrinology 130, 1707-1715. MEDLINE

Chen TC, Hinton DR, Zidovetzki R and Hofman FM. (1998). Lab. Invest. 78, 165-174. MEDLINE

D'Angelo G, Lee H and Weiner RI. (1997). J. Cell. Biochem. 67, 353-366. MEDLINE

Dolci S, Williams DE, Ernst MK, Resnick JL, Brannan CI, Lock LF, Lyman SD, Boswell HS and Donovan PJ. (1991). Nature 352, 809-811. MEDLINE

Godin I, Deed R, Cooke J, Zsebo K, Dexter M and Wylie CC. (1991). Nature 352, 807-809. MEDLINE

Godwin AK, Perez RP, Johnson SW, Hamaguchi K and Hamilton TC. (1992). Hematol. Oncol. Clin. North Am. 6, 829-841. MEDLINE

Heinrich MC, Dooley DC and Keeble WW. (1995). Blood 85, 1769-1780. MEDLINE

Horie K, Takakura K, Taii S, Narimoto K, Noda Y, Nishikawa S, Nakayama H, Fujita J and Mori T. (1991). Biol. Reprod. 45, 547-552. MEDLINE

Huang EJ, Nocka KH, Buck J and Besmer P. (1992). Mol. Biol. Cell 3, 349-362. MEDLINE

Ismail RS, Dubé M and Vanderhyden BC. (1997). Dev. Biol. 184, 333-342. MEDLINE

Ismail RS, Okawara Y, Fryer JN and Vanderhyden BC. (1996). Mol. Reprod. Dev. 43, 458-469. MEDLINE

Jindal SK, Snoey DM, Lobb DK and Dorrington JH. (1994). Gyn. Oncol. 53, 17-23.

Julian CG. (1974). Clin. Obstet. Gynecol. 17, 241-257. MEDLINE

Karlan BY, Bladwin RL, Cirisano FD, Mamula PW, Jones J and Lagasse LD. (1995). Gynecol. Oncol. 59, 67-74. MEDLINE

Kruk PA and Auersperg N. (1994). In Vitro Cell. Dev. Biol. Anim. 30A, 217-225. MEDLINE

Lammie A, Drobnjak M, Gerald W, Saad A, Cote R and Cordon-Cardo C. (1994). J. Histochem. Cytochem. 42, 1417-1425. MEDLINE

Lidor YJ, Xu FJ, Martinez-Maza O, Olt GJ, Marks JR, Berchuck A, Ramakrishnan S, Berek JS and Bast Jr RC. (1993). Exp. Cell Res. 207, 332-339. MEDLINE

Manova K, Huang EJ, Angeles M, De Leon V, Sanchez S, Pronovost SM, Besmer P and Bachvarova RF. (1993). Dev. Biol. 157, 85-99. Article MEDLINE

Matsui Y, Toksoz D, Nishikawa S, Nishikawa SI, Williams D, Zsebo K and Hogan BLM. (1991). Nature 353, 750-752. MEDLINE

Mekori YA and Metcalfe DD. (1994). J. Immunol. 153, 2194-2203. MEDLINE

Metcalf D and Nicola NA. (1991). Proc. Natl. Acad. Sci. USA 88, 6239-6243. MEDLINE

Motro B and Bernstein A. (1993). Dev. Dynamics 197, 69-79.

Murdoch WJ. (1996). Biol. Rev. Camb. Philos. Soc. 71, 529-543. MEDLINE

Murphy ED. (1972). J. Nat. Cancer Inst. 48, 1283-1295.

Nicosia SV, Johnson JH and Streibel EJ. (1985). Int. J. Gynecol. Pathol. 4, 58-74. MEDLINE

Nishina Y, Kobarai Y, Sumi T, Kosaka M, Nishikawa S and Nishimune Y. (1992). Exp. Cell Res. 198, 352-356. MEDLINE

Nocka K, Tan JC, Chiu E, Chu TY, Ray P, Traktman P and Besmer P. (1990). EMBO J. 9, 1805-1813. MEDLINE

Ogawa K, Takeda Y, Tashima M, Sawai H, Toi T, Okazaki T, Sawada H, Maruyama Y and Okuma M. (1995). Blood 85, 1496-1503. MEDLINE

Osterholzer HO, Streibel EJ and Nicosia SV. (1985). Biol. Reprod. 33, 247-258. MEDLINE

Packer AI, Hsu YC, Besmer P and Bachvarova RF. (1994). Dev. Biol. 161, 194-205. MEDLINE

Parrott JA and Skinner MK. (1997). Endocrinology 138, 3819-3827. MEDLINE

Pesce M, Farrace MG, Piacentini M, Dolci S and De Felici M. (1993). Development 118, 1089-1094. MEDLINE

Ramsfjell V, Borge OJ, Cui L and Jacobsen SE. (1997). J. Immunol. 158, 5169-5177. MEDLINE

Reith AD, Rottapel R, Giddens E, Brady C, Forrester L and Bernstein A. (1990). Genes Dev. 4, 390-400. MEDLINE

Rodriquez GC, Berchuck A, Whitaker RS, Schlossman D, Clarke-Pearson DL and Bast Jr RC. (1991). Am. J. Obstet. Gynecol. 164, 745-750. MEDLINE

Rossi P, Albanesi C, Grimaldi P and Geremia R. (1991). Biochem. Biophys. Res. Commun. 176, 910-914. MEDLINE

Rossi P, Dolci S, Albanesi C, Grimaldi P, Ricca R and Geremia R. (1993). Dev. Biol. 155, 68-74. Article MEDLINE

Sambrook J, Fritsch EF and Maniatis T (eds). (1989). Molecular Cloning, a Laboratory manual. Cold Spring Harbor Laboratory: Cold Spring Harbor, New York.,

Skinner MK, Keski-Oja J, Osteen KG and Moses HL. (1987). Endocrinology 121, 786-792. MEDLINE

Tisdall DJ, Quirke LD, Smith P and McNatty KP. (1997). J. Mol. Endocrinol. 18, 127-135. MEDLINE

Tortora G, di Isernia G, Sandomenico C, Bianco R, Pomatico G, Pepe S, Bianco AR and Ciardiello F. (1997). Cancer Res. 57, 5107-5111. MEDLINE

Tsai M, Takeishi T, Thompson H, Langley KE, Zsebo KM, Metcalfe DD, Geissler EN and Galli SJ. (1991). Proc. Natl. Acad. Sci. USA 88, 6382-6386. MEDLINE

Vadiveloo PK, Vairo G, Novak U, Royston AK, Whitty G, Filonzi EL and Cragoe Jr EJ and Hamilton JA. (1996). Oncogene 13, 599-608. MEDLINE

Vigne JL, Halburnt LL and Skinner MK. (1994). Biol. Reprod. 51, 1213-1221. MEDLINE

Zheng W, Magid MS, Kramer EE and Chen YT. (1996). Am. J. Pathol. 148, 47-53. MEDLINE

Ziltener HJ, Maines-Bandiera S, Schrader JW and Auersperg N. (1993). Biol. Reprod. 49, 635-641. MEDLINE

Figure 1 Localization of KL mRNA in the rat ovarian surface epithelium by in situ hybridization. Photographs are brightfield images of 6 m sections through ovaries of 28-day-old PMSG-treated rats. Sections are hybridized with a KL probe in the absence (a, b and c) or presence of 100-fold excess unlabelled KL probe (d). A region of the section shown in (b) is magnified in (c)

Figure 2 Immunohistochemical detection of KL protein in sections of ovaries from 28-day-old PMSG-treated rats. Sections were incubated with antibodies against KL and a diffuse brown product indicating positive KL staining is seen in both surface epithelial cells and in granulosa cells (a). In (b), immunohistochemistry was performed without the primary KL antibodies to identify non-specific staining

Figure 3 Northern blot analysis of Kit receptor and KL mRNA expression in primary cultures of rat ovarian surface epithelial (OSE) cells and immortalized OSE cells (ROSE 199). Blots with total RNA were sequentially probed with ³²P-labelled cDNA probes for KL (a), Kit receptor (b) and -tubulin (c). NIH3T3 fibroblasts and MC/9 mast cells served as KL positive/Kit negative and KL negative/Kit positive controls, respectively

Figure 4 Reverse transcription-polymerase chain reaction was performed on total RNA from rat OSE cells and ROSE 199 cells. KL primers spanning exon 6 of KL mRNA were used to amplify KL-1 (287 bp) and KL-2 (203) transcripts. C13, a KL-producing epithelial cell line, and granulosa cells were used as positive controls. RNA was substituted with water prior to RT - PCR to serve as a negative control. In some experiments, MC/9 cells were used as negative control cells (data not shown)

Figure 5 Northern blot analysis of KL expression in ROSE 199 cells cultured in the presence or absence of 10 ng/ml TGF- and/or 1 mM dbcAMP. Blots were probed sequentially with ³²P-labelled cDNA probes for rat KL (a) and -tubulin (b). Total RNA samples from NIH3T3 and Sl ⁴ fibroblasts served as positive and negative KL mRNA controls, respectively. Band intensity was estimated using ImageQuaNT software and the ratios of KL/-tubulin values were standardized to control levels which were arbitrarily set to 1.0 (c). Bars represent the mean±s.e.m. values of three experiments. Those indicated with an asterisk are significantly different from control (P<0.05)

Figure 6 Reverse transcription-polymerase chain reaction performed on ROSE 199 cells cultured in the presence or absence of 10 ng/ml TGF- and/or 1 mM dbcAMP. NIH3T3 fibroblasts and MC/9 mast cells served as KL positive/Kit negative and KL negative/Kit positive controls, respectively. KL primers were used to amplify the KL-2 (203 bp) and KL-1 (287 bp) transcripts. RNA was substituted with water prior to RT - PCR to serve as a negative control

Figure 7 Proliferation of ROSE 199 cells cultured for 48 h in the presence of 1 mM dbcAMP and/or 10 ng/ml TGF-. Each data point represents the mean±s.e.m. of at least four experiments with duplicate treatments in each experiment. Bars with different letters are significantly different (P<0.05)